Introduction

Prospective memory (PM) is the ability to remember to carry out planned events in a specific future situation (McDaniel and Scullin, 2010). In the past several decades, most studies in the PM field have been interested in the forgetting of PM tasks and the causes of PM omission errors (e.g., failure to implement the PM intention). Recently, PM research has begun to pay attention to the activation state of PM intention and its aftereffects after PM task completion. PM aftereffects refer to the phenomenon that participants retain the PM intention after the PM task is completed or canceled, which may interfere with subsequent task performance or cause commission errors by performing previous PM tasks repeatedly when they are no longer relevant (Matos et al., 2020). In the current study, we mainly focused on commission errors generated by completed, but no longer relevant PM intentions. Commission errors of PM can have serious consequences. For example, an older person who takes medicine repeatedly may experience adverse outcomes including life-threatening outcomes. A large body of research has also found the existence of commission errors of PM in different age groups (Cottini and Meier, 2020; Meier and Rey-Mermet, 2012). Furthermore, when PM intentions continue to remain active after completion, they occupy attention resources and interfere with the execution of other tasks (Matos and Albuquerque, 2021; Meier and Rey-Mermet, 2018).

To date, most studies have adopted event-based PM paradigms with no-longer-relevant PM cues to assess commission errors of completed intentions with two phases: a PM active phase and a PM finished phase (Möschl et al., 2020; Scullin et al., 2012). During the PM active phase, participants are usually asked to execute both an ongoing task and an embedded PM task simultaneously, after which they are explicitly informed that the PM task has been completed or canceled. During the subsequent PM finished phase, participants only need to perform the ongoing task, but the PM cues that have been completed in the previous phase will reappear as ongoing stimuli (Bugg et al., 2016; Matos et al., 2020; Möschl et al., 2020). Such a procedure can measure the specific situation of a PM task before and after it is completed. However, in reality, an intention is seldom executed in isolation, rather multiple PM tasks are constantly being formed, maintained, retrieved, and performed. For example, remember to turn off the gas stove after cooking, take medicine after meals, and bring the previously written work report before going out. The new PM intention often occurs after the completion of the old PM intention, and the formation of a new PM intention is likely to have an impact on the old PM intention.

Several theoretical viewpoints can explain the effect of new PM intentions on PM commission errors. First, the attentional dependence hypothesis posits that a new PM task does not change the activation level of completed intentions, but rather triggers a shift in attention allocation. Forming a new intention will pre-activate the currently relevant PM response, and increase the state of readiness for attention to the new PM response (Walser et al., 2017). In addition, forming a new PM task will establish a general and continuous PM task set that includes both new and old PM tasks. This PM task set allows individuals to continuously monitor for PM cues among ongoing task stimuli, thus preventing them from completely ignoring the emergence of novel stimuli. When participants notice new PM cues, previous cues that are no longer relevant will also be accidentally retrieved. In particular, when old and new PM tasks are similar, old PM intentions will be erroneously retrieved as a new PM task, thus promoting the occurrence of commission errors (Walser et al., 2014). According to the attentional dependence hypothesis, completing a new PM task in the PM finished phase will increase the probability of an individual making PM commission errors. By comparison, the intention overwriting hypothesis claims that forming a new intention will overwrite the components of the completed intention representation, including PM cues, intended actions, and PM cue-intended action associations, thus changing the activation level of the previously completed intention representation (Walser et al., 2012). In particular, changes in PM tasks may lead to a decrease in the readiness of the completed intention components, and destroy the link between previous PM cues and intentions, which will reduce an individual’s attention to PM cues and the possibility of spontaneous PM intention retrieval. Thus, the execution of new PM tasks during the finished phase reduces an individual’s PM commission errors (Smith et al., 2007).

Several studies have directly focused on the impact of executing new PM tasks on PM commission errors. For example, Walser et al. (2012) investigated the commission errors of PM in the repeated PM cue paradigm. They set up 8 blocks in Experiment 1, and PM cues were different for each block. Compared to the previous block, the latter block was equivalent to updating a new PM task. However, participants were asked to perform the ongoing task only (equivalent to no new PM tasks added) in even-numbered blocks in Experiment 3. By comparing the difference in response times to previous PM cues (repeated PM cue trials) and control trials (oddball trials) in Experiment 1 and Experiment 3, it was found that the response time’s difference in Experiment 1 significantly increased more than in Experiment 3. Moreover, the PM commission error rate in Experiment 1 was higher than that in Experiment 3, which indicated that the aftereffect of the completed intention was significantly smaller in ongoing-task-only conditions than that in new-PM-task conditions (Experiment 1). These results are consistent with the attentional dependence hypothesis. Walser et al. (2017) adopted a modified no-longer-relevant repeated PM cue paradigm in Experiment 3. The finished phase of this paradigm included a PM-task-repetition condition and a PM-task-switch condition. They further compared the ongoing-task-only condition (no new PM tasks were added in the finished phase), PM-task-repetition conditions (a new PM task was added in the finished phase, and its cue and response characteristics matched those of the previous PM task), and a PM-task-switch condition (a new PM task was added in the finished phase, and its cue and response characteristics mismatched those of the previous PM task). Results showed that only participants in the PM-task-repetition condition made more PM commission errors, while participants in the ongoing-task-only condition and PM-task-switch condition made very few PM commission errors. The results partially supported the attentional dependence hypothesis. However, Anderson and Einstein (2017) investigated the impact of forming a new PM task on commission errors of PM intentions in Experiment 2. They found that the commission error rate of the new PM intention group was not different from that of the control group, which was consistent with the intention overwriting hypothesis. The difference between the results of the two studies may be due to the difference in similarity between new and old PM cues. The old and new PM cues set by Walser et al. all adopted rectangles with different characteristics, and the new cues were similar to the old ones. Similar PM cues are easily confused, and individuals are likely to mistake old cues for new ones, resulting in more commission errors. However, the old and new PM cues set by Anderson and Einstein used words with great semantic differences. Individuals’ memory of new words will inhibit the memory of original words (Tulving and Psotka, 1971). So when new PM cues are presented, the memory of old PM cues may be inhibited and the commission error rate may be reduced.

Implementation intention is an encoding method that closely links the expected situation with the intended behavior. This strategy involves encoding tasks in the form of an “if-then” statement (if situation X occurs, then I will perform intention Y) (Chen et al., 2019; Gollwitzer and Sheeran, 2006; Kretschmer-Trendowicz et al., 2021). As an effective PM encoding strategy, implementation intention has been shown to increase the association between cues and intended actions and significantly improve PM performance (Bugg et al., 2013). In a meta-analysis, Chen et al. (2015) included 41 studies focusing on the impact of implementation intention on PM, of which 34 included additional mental visualization of the intention execution, while other studies adopted additional strategies like repetition. Most of those studies did not simply use the “if-then” encoding method, but rather the method was often accompanied by other additional encoding strategies, which not only promoted the association between a PM cue and action but also improved the activation level of intention, thus promoting PM performance (Chen et al., 2015; Henry et al., 2020; Webb and Sheeran, 2008).

Implementation intention has also been found to significantly influence PM commission errors, and several studies have shown that implementation intention can increase the risk of commission errors in an active PM task (i.e., active PM intentions). For example, Meiser and Rummel (2012), compared the effect of an implementation intention condition (the form of “if-then” plus imagination), mixed conditions (implementation intention plus practice), and control conditions on PM commission errors in young people. The results showed that, compared with the control condition, commission errors were higher in the implementation intention and mixed conditions. The researchers believed that the implementation intention encoding enhanced the spontaneous retrieval of PM cues. Moreover, the occurrence of commission errors also benefited from the spontaneous retrieval of PM cues during the PM active phase. Bugg et al. (2013) have also found that implementation intention similarly promoted PM commission errors in older age groups in the completed PM paradigm (implementation intention encoding including the form of “if-then” + imagination + retelling).

If we adopt an implementation intention to strengthen the new PM intention in the completion phase, will this affect the PM commission errors to a greater extent? The current study mainly focused on the effects of performing new PM tasks on commission errors through the implementation intention encoding strategy in the finished phase. The strong associations between PM cues and actions increase the likelihood of spontaneous retrieval of completed PM intentions, facilitating the occurrence of commission errors (Scullin and Bugg, 2013). The implementation intention in the active phase promotes the association between PM intentions and behavior (Chen et al., 2015), and this effect extends to the finished phase, which encourages individuals to commit more commission errors. However, using the implementation intention to encode a new PM task in the finished phase has an uncertain effect on previous PM intentions. According to the intention overwriting hypothesis, using implementation intention to encode a new PM intention in the finished phase should overwrite the previous old intention more effectively, thus reducing activation of the old intention, and reducing commission errors. However, the attentional dependence hypothesis suggests that the implementation intention will promote the allocation of attention resources from the ongoing task to new PM tasks. This results in increased accidental retrieval of the old intention related to the new intention, which is reflected in increased commission errors and poorer ongoing task performance. By manipulating implementation intention encoding, the roles of the intention overwriting hypothesis and attentional dependence hypothesis in explaining the impact of performing new tasks on commission errors can be clarified. In addition, we can also confirm whether implementation intention is an effective method for reducing commission errors (if the intention overwriting hypothesis is supported, then it shows that implementation intention can effectively reduce commission errors). Therefore, the second aim of our study was to examine the coverage effect of new intentions formed by the implementation intention encoding strategy on the completed intention in the PM finished phase.

The present study

This study explored the effect of forming new PM intentions on PM commission errors by comparing the differences between a non-new PM group (without adding PM tasks), a new PM group (adding new PM tasks), and an implementation intention group (adding PM tasks using implementation intention encoding) in the finished phase. The attentional dependence hypothesis suggests that new PM tasks will cause individuals to form a PM task set and incorrectly retrieve old PM intentions, resulting in more PM commission errors. If the commission error rates of the three groups increase sequentially (non-new PM group < new PM group < implementation intention group), then the attentional dependence hypothesis is supported. In contrast, the intention overwriting hypothesis claims that the new PM intention will overwrite the components of the old PM intention, and reduce the preparatory attention state of the completed PM intention. The strength of the association between the previous PM cue and the intended action will also be reduced, thereby resulting in fewer commission errors. According to the intention overwriting hypothesis, if the commission error rates of the three groups decrease sequentially (non-new PM group > new PM group > implementation intention group), the intention overwriting hypothesis is supported.

Methodology

Participants

We used G*Power (Effect size f = 0.4, α = 0.05, 1 − β = 0.8) to test the sampling size, and the lowest number of participants was 66. In order to achieve the minimum number of participants who made commission errors, we recruited participants until the number of participants making commission errors reached 66. Finally, two hundred and thirteen college students (87 males; Mage = 20.17, SD = 1.36; range = 18–25) were recruited to participate in the study and were tested individually, of which 66 made commission errors. All participants signed an informed consent form before commencing the experiment and received 15 yuan as a reward upon completion. This study was jointly approved on 8 May 2021 by three members of the Academic Committee of Henan Provincial Key Laboratory of Psychology and Behavior, one of whom is the chair of the Academic Committee. The ethical approval number is 20210508017.

Experimental design

We employed a single factor between-subjects design that included three groups: a non-new PM group, a new PM group, and an implementation intention group. Participants were randomly assigned to the new PM group (N = 71), the non-new PM group (N = 71), or the implementation intention group (N = 71).

Experimental procedure

We adopted E-prime 2.0 to write the experimental procedure. The materials adopted 24 English letters except for F and J. The size of each letter was about 2 cm×2 cm, and the viewing angle was about 1.43°. The PM cues were the G and Q (or R and W), and the remaining 22 letters were used as 1-back task stimuli.

At the beginning of the procedure, the 1-back task instruction was presented first. Participants were told that they needed to compare two adjacent letters. If the two letters were the same, they were to press the J key, otherwise, they were to press the F key (the F key and J key were labeled “different” and “same” respectively). After that, they were required to practice 50 1-back tasks (Guo and Gan, 2022). A plus fixation appeared for 500 ms at the beginning of each trial followed by the presentation of an English letter for no more than 2 s which disappeared after participants responded. Finally, an empty screen appeared for 500 ms. Only participants who achieved a correct rate of more than 60% in the practice stage were permitted to participate in the formal experiment.

The formal experiment consisted of a PM active phase (phase 1) followed by a PM finished phase (phase 2) (Möschl et al., 2020). The PM active phase commenced with the PM task instructions: you should press the space bar when the letters G or R appear while performing the ongoing task. Before the task, participants were asked to press the F key, J key, and space bar in turn to become familiar with the key responses. The PM cues appeared four times, in which the letters G and R appeared two times each, and the phase comprised 160 trials of ongoing tasks. Following this, the PM finished phase commenced and participants in all groups were initially informed that the PM task of phase 1 was completed and no longer needed to be performed. In phase 2, participants in the non-new PM group were only required to perform the ongoing task, in which the completed PM cues G and R still appeared. Participants in the new PM group were told to press the space bar when they encountered the letters Q or W.

However, the completed PM cues G and R appeared as the trial of the ongoing task. The implementation intention group was similar to the new PM group except that the new PM task was presented in the form of “if-then” as follows: if the letters Q or W were encountered, then the space bar should be pressed directly without executing the letter comparison task (using the “if-then” encoding form), and participants were asked to repeat the PM task aloud three times (Zimmermann and Meier, 2010). There were 320 trials of ongoing tasks in the PM finished phase, in which eight old PM cues were inserted, four for each of G and R. The 22 letters that made up the ongoing task stimulus appeared in a pseudo-random manner during the procedure, with each letter appearing with the same frequency. Throughout the procedure, the ratio of the same and different responses that participants were required to perform was 1:2. Also, all PM cues did not appear twice in a row. In the new PM tasks, new PM cues appeared eight times, with Q and W appearing four times. After the experiment concluded, participants were requested to articulate the content of the PM task (both old and new) to confirm that they remembered it (Gan and Guo, 2019). The flowchart of experimental stimuli is shown in Fig. 1.

Fig. 1: Experimental procedures for active phase and finished phase.
figure 1

When each letter appears, participants need to respond by pressing the F key, J key, or spacebar.

Full size image

Results

Commission errors were defined as a participant pressing the space key when the old completed PM cues appeared during the PM finished phase (Scullin et al. 2012). Analysis of Variance (ANOVA) was used to analyze differences between the groups for continuous outcomes, and chi-square test was performed to assess group differences in the case of categorical outcome variables. Analyses were performed using SPSS 20.0.

Results of the chi-square test showed that the proportion of participants who committed at least one commission error did not differ across the three groups: the non-new PM group (number and proportion of commission errors: 31 out of 71), the new PM group (20 out of 71) and the implementation intention group (15 out of 71), χ2(2, N = 213) = 8.83, p < 0.05, φ = 0.21. Multiple comparisons using Bonferroni showed no difference between the non-new PM group and the new PM group, p > 0.05. There was also no difference between the new PM group and the implementation intention group, p > 0.05. The number of commission errors in the non-new PM group was significantly higher than that in the implementation intention group, χ2(1, N = 142) = 8.36, p < 0.01, φ = 0.24. We excluded the data of participants who did not make a commission error and further compared the number of different groups who made only one commission error using the chi-square test and found no difference among the three groups, p > 0.05. Among the number of people who made commission errors, 13% of the non-new PM group made one commission error, 20% of the new PM group made one commission error, and 40% of the implementation intention group made one commission error.

Commission errors

We excluded data from participants who did not make commission errors, and ultimately analyzed the commission error rate of participants who made commission errors with ANOVA. The results showed that commission errors differed significantly among the three groups, F(2, 63) = 8.47, p = 0.001, ηp2 = 0.21. Multiple comparisons revealed that the new PM group had a lower commission error rate than the non-new PM group (p < 0.01), indicating that performing new PM tasks in the finished phase reduced commission errors. Besides, the implementation intention group had a significantly lower commission error rate than the non-new PM group (p < 0.001), indicating that the implementation intention encoding further decreased commission errors (see Fig. 2).

Fig. 2: The rate of commission errors in the finished phase.
figure 2

Two asterisks represent p < 0.01. The error bar is a standard error.

Full size image

Moreover, the results of the ANOVA showed that the differences among the three groups did not reach significance in their reaction time to the old PM cues in the finished phase, p > 0.05.

Ongoing task performance

A one-way ANOVA was conducted for the accuracy of ongoing tasks, and results showed that the differences among the three groups were not significant, p > 0.05. A one-way ANOVA was conducted for the reaction time of ongoing tasks, and results showed that the differences among the three groups were not significant, p > 0.05 (see Fig. 3).

Fig. 3: The accuracy of ongoing tasks in the finished phase.
figure 3

The error bar is a standard error.

Full size image

PM task performance

Since new PM tasks were only presented in the new PM group and the implementation intention group in the PM finished phase, we only compared new PM task performance between those two groups. Results of an ANOVA revealed a significant difference between the two groups, F (1, 142) = 23.64, p < 0.001, ηp2 = 0.14. The paired comparisons revealed that the implementation intention group was more accurate than the new PM group, indicating that the implementation intention encoding effectively improved the performance of the new PM task. There was, however, no difference between the two groups in their reaction speed to the new PM task, p > 0.05 (Table 1).

Table 1 The performance of multiple tasks in the PM finished phase (M ± SD).
Full size table

Discussion

The intention overwriting hypothesis and the attentional dependence hypothesis make different predictions about how new PM tasks affect PM commission errors. In terms of commission errors, the intention overwriting hypothesis predicts that a new PM intention will decrease an individual’s PM commission errors, while the attentional dependence hypothesis predicts that the individual will have an increased probability of making PM commission errors. Compared to the non-new PM group, the implementation intention group had fewer participants who made commission errors, indicating that strengthened new PM intention reduces the number of individuals making commission errors. In addition, the commission error rate of the new PM group and implementation intention group is lower than that of the non-new PM group, indicating that adding new PM tasks reduced the degree of commission errors, supporting the intention overwriting hypothesis. However, this result is inconsistent with some studies. For example, Walser et al. (2012) found that the new PM intention increased the commission error rate, which may be because the old and new PM cues they set adopted rectangles with different characteristics, and the new cues were similar to the old ones. Similar PM cues are easily confused, and individuals are likely to mistake old cues for new cues, resulting in more commission errors. It is worth noting that compared to the non-new PM group, the number of people who make commission errors in the new PM group has not decreased, but the rate of commission errors has decreased. This indicates that adding only one ordinary PM intention may not initially effectively suppress the old PM intention, so there are still many people who make mistakes. However, with the successive execution of the new PM task, the degree of activation of the new PM task is gradually increased, and the suppression of the old PM intention is gradually strengthened, resulting in the subsequent PM task becoming less and less prone to error. Therefore, the new PM group will have a phenomenon in which the number of participants making mistakes does not decrease, but the error rate decreases. The new PM intention of the implementation intention group is strengthened by the implementation intention coding at the beginning, which can effectively suppress the original PM intention. So the number of participants who make commission errors and the error rate of the implementation intention group are significantly reduced. In addition, the experimental paradigm adopted by Anderson and Einstein (2017) was consistent with this study, but they did not find significant differences between the new PM group and the control group, which may be because the PM cues they used were non-focal cues. The recognition of focal cues has been included in the process of ongoing tasks and is easy to extract spontaneously, while the recognition of non-focal cues is quite different from the process of ongoing tasks and requires extra attention (Zuber et al., 2016). Therefore, the non-focal cues of old PM tasks are not easy to extract after completion, and the degree of inhibition of new PM intentions on the old PM intention may be weak. In this study, focal cues are used. Old PM cues are easy to extract spontaneously, and new PM intentions have a larger space to inhibit old PM intentions, so the effect is more obvious.

The old and new PM cues used in the two studies by Walser et al. (2012, 2017) were salient and similar symbols. The perceptual similarity between old and new PM cue features can easily be confused, which may lead an individual to commit more commission errors. By comparison, the old and new PM task cues used in our experiment were two clear and distinct English letters, and hence participants were less likely to confuse the old and new PM tasks. In addition, Walser et al. (2017) changed the similarity between the old and new PM tasks in the PM-task switch condition but found no difference in commission error rates between the PM-task-switch condition and the ongoing-task-only condition, which to some extent supports our speculation. One reason for the lack of difference in the error rate between the two conditions may be the occurrence of a floor effect, with a somewhat lower error rate in the PM-task-switch condition compared to the ongoing-task-only condition. This finding was also supported in the study of Anderson and Einstein (2017). They also found no significant difference in commission errors among different conditions in the PM cue paradigm, although the commission error rate in the new PM task condition tended to be lower than that of the control condition. Therefore, the results of our study may not conflict with those of Walser et al. (2017) and Anderson and Einstein (2017). Notably, although we found that the new PM group had a lower probability of making commission errors than the non-new PM group, there was no difference in the proportion of participants who made at least one commission error. Further analysis found that the probability of making two or more mistakes in the new PM group was significantly lower than that in the non-new PM group, indicating that the addition of new PM tasks did not lead to fewer participants committing commission errors, but rather reduced the degree to which they made mistakes.

How does the new PM task reduce PM commission errors? According to the intention overwriting hypothesis, the new PM intention overwrites the completed PM intention, including its old PM cue, intended action, and the association between cue and intended action, thereby reducing the activation level of completed intention representations (Walser et al., 2017). The results of the implementation intention condition in our study also support the intention overwriting hypothesis. Most of the implementation intention encoding adopted by most current studies will add additional repetition or use of imagination to the “if-then” encoding form, which not only strengthens the association between PM cues and intended actions but also further improves the activation level of PM intentions. (Chen et al., 2015; Gollwitzer and Sheeran, 2006). The results of this study found that the implementation intention group had a lower commission error rate in the finished phase, indicating that the enhancement of new PM intentions further reduced the possibility of retrieval of old PM intentions, which is consistent with the intention overwriting hypothesis. In a sense, the coverage of old PM intentions by new PM intentions is actually an interference effect.

Walser et al. (2014) found that the aftereffects of completed tasks were reduced when a filling task with a cognitively high-demanding load was performed between the active phase and the finished phase. They argued that the deactivation of a completed intention is a function of new information replacing or interfering with memory representations of old intentions and that any task requiring cognitive demand helps to overwrite representations of previous intentions (especially the association between PM cues and intended actions). Consistent with this inference, old completed PM intentions should have been overwritten or interfered with when performing new PM tasks that require cognitive resource allocation during the finished phase in the current study. Previous studies have demonstrated that changes in PM response types do not affect PM commission errors (Walser et al., 2017), although the new PM task we added retained the same key response as the previous PM task, the new cues were changed so they were not easily confused with the old PM cues. Such manipulation may shake or even disintegrate representations of old intentions. Furthermore, we adopted the implementation intention to encode the new PM task, which promoted the automatic association between new PM cues and intended actions (Chen et al., 2015). These new intentions, efficiently encoded by the implementation intention strategies, may also accelerate the destruction of the association between old PM cues and intended actions.

From the perspective of retrospective memory, we suspect that the process of overwriting old intentions with new intentions may be a process of retrospective interference or a process of retrieval-induced forgetting. Studies on retrieval-induced forgetting suggest that individuals retrieving part of the information from long-term memory will passively suppress other relevant information, making it easier to forget this information (Anderson et al., 2000). If the new and old PM intentions are regarded as a task set, then when new PM intentions are formed and retrieved, the contents of the old PM intentions are naturally suppressed, and forgetting is accelerated which reduces the possibility of retrieving old PM intentions. Moreover, the activation level of old PM intentions decreases over time (Walser et al., 2012), while new PM intentions are at a higher activation level due to their short formation time in memory, making new PM intentions easier to retrieve. In this study, we also found that compared with the non-new PM group, the number of commission errors made by the new PM group did not decrease, but the degree of errors was reduced, possibly because participants forgot old PM intentions more rapidly when they initially retrieved new PM intentions (we also found that participants were more likely to make commission errors when the previous PM cues first appeared). However, Walser et al. (2017) found that performing new PM tasks in the PM-task-repetition condition increased the PM commission error rate, which may be because the old and new cues they used had salient and similar characteristics, leading to confusion about the two PM tasks. Since new PM intentions were highly activated, old PM intentions would be more likely to be incorrectly noticed and retrieved. Therefore, depending on the similarity of the new and old PM intentions, there may be two different mechanisms underpinning the impact of new PM intentions on old PM intentions. From this perspective, the attentional dependence hypothesis is reasonable to some extent, but it can only explain the results when the old and new PM tasks are relatively similar.

Both the attentional dependence hypothesis and the intention overwriting hypothesis predict changes in attention caused by the formation of new PM intentions. Our study found no difference between the non-new PM task group, the new PM task group, and the implementation intention group in the accuracy and response time of the ongoing task, as well as the response time to the old PM cue during the finished phase. This indicates that the execution of new PM tasks was not accompanied by additional attention consumption, nor did it create attention changes in old PM tasks. This is not only inconsistent with the attentional dependence hypothesis and the intention overwriting hypothesis but also differs to previous research findings. In our study, the ongoing tasks were mainly used to indirectly reflect the attention changes of the old and new PM tasks. However, the focality of PM task cues (i.e. PM cues may be considered focal or non-focal depending on whether or not target properties associated with the intention are in the focus of attention during ongoing task performance) can have an impact on ongoing task performance (Einstein et al., 2005). PM tasks with focal cues are less dependent on attention, while PM tasks with non-focal cues usually require more attention (Scullin et al., 2010). By comparison with Walser et al. (2012, 2017) who used non-focal cues, all cues used in the present study were focal cues, which may have resulted in the new PM task automatically capturing more attention. This is one possible reason for the difference between our study results and those of previous studies. Second, the 1-back task used in our study was relatively simple, so it was difficult to ensure that the individual’s attention was fully occupied by the ongoing task. In this case, an individual is likely to have had sufficient attention remaining to process the new PM task, and the performance of the ongoing task may not have sensitively reflected the attention consumption demanded by the new PM task. In conclusion, due to the characteristics of the new PM task and the limitations of the ongoing task, we cannot confidently draw convincing conclusions regarding the indicators of attention change.

This study explored the effects and processing mechanisms underpinning PM commission errors when executing a new PM task in the finished phase. The results showed that adding new PM intentions significantly reduced the probability of individuals making commission errors. Moreover, when the new PM task was encoded using the implementation intention strategy, the probability of individuals committing commission errors decreased further. It is worth noting that the attention dependence hypothesis and intention overwriting hypothesis are both based on conjectures with limited evidence and have not received sufficient evidence to verify them, nor have they predicted a null effect. For example, existing studies have not added delays between the active and finished stages of the original PM task, while in real life, there are usually time intervals between multiple PM tasks. Delaying can reduce the activation level of PM tasks (McBride et al., 2011), and individuals may have weaker original PM intentions under longer delay conditions. The inhibitory effect of new PM intentions on old PM intentions may not be significant, and a null effect may occur. At present, neither hypothesis can predict this phenomenon.

Conclusions

In everyday life, we often form multiple PM intentions, and the formation of new PM intentions is likely to have an impact on old ones. This study focused on the effect of the formation of new PM intentions on commission errors considering the intention overwriting hypothesis and attention dependence hypothesis. The results showed that executing a new PM task during the PM finished phase significantly reduced the commission error rate compared to the non-new PM task group. Furthermore, the adoption of implementation intention strengthened the encoding intensity of the new PM task, and we found that commission errors were further reduced. Therefore, the overall results of this study revealed that the formation of a new PM intention may effectively reduce PM commission errors.